Browse documentation

KPI lenses & lift potential

Recolor the landscape by any KPI to see where each metric concentrates, and read how segments over- or under-index against the population.

The segment landscape starts as a map of population. KPI lenses add the dimension you actually care about: pick any KPI from your use case definition and the landscape recolors to show how that metric distributes inside every segment, so you can see at a glance which groups concentrate the behavior you want to move.

Switching KPI lenses

The KPI bar sits at the top of the Discovery tab, with one pill per KPI defined in the use case definition.

  • Hover a pill to see that KPI’s definition.
  • Click a pill to activate the lens.
  • Click the × on the active pill to clear it and return to the plain population view.

Only one lens is active at a time. Each lens is a different question asked of the same landscape.

Reading a lens

With a lens active, every segment tile gains a proportional colored bar showing the breakdown of that KPI’s values within the segment. Colors are consistent across the whole landscape, so the same KPI value is the same color in every tile. Scan for a color and you’ve found the segments where that value concentrates.

The run metrics bar gains a KPI Distribution chart showing the same breakdown at the population level. That’s your baseline for reading lift: a segment over-indexes on a KPI value when its own distribution skews toward that value more than the population does, and under-indexes when it skews away. Select a segment and the bar shows the run-level and segment-level distributions side by side, making the comparison direct. No mental math required.

Over- and under-indexing is what turns a segment into a target audience. A segment that over-indexes on churn risk is a retention audience; one that under-indexes on a product category is a cross-sell audience. The segment tactics behind the run’s actions build on exactly these contrasts.

What-if lift modeling (Preview)

For KPIs flagged with lift potential in the definition, the Discovery tab can go one step further and model a hypothetical scenario. This capability is currently in Preview.

When you activate a lift-enabled lens, a Lift Potential toggle appears below the KPI bar. Switching it on replaces the population view with a value view:

  1. A scenario form appears, reading If [X] % of Customers were lifted by [Y], followed by the KPI’s own unit. Enter the share of customers you believe a campaign could move, and by how much.
  2. The metrics bar switches from Customers to Total Lift Potential, an estimated dollar value for the scenario across the run; hover it for the unrounded figure.
  3. The distribution chart becomes KPI Lift Distribution, showing where that value sits across KPI values.
  4. Select a segment to see its own Segment Lift Potential, so you can compare which segments carry the most headroom under your scenario.

Adjust the two numbers and the estimates update. Use it to pressure-test how ambitious a campaign needs to be before the value justifies the effort.

Note. Lift potential figures are scenario estimates, not forecasts. They tell you where value is concentrated if your assumptions hold; they don’t promise a campaign outcome.

Next steps